https://github.com/mariusae/heapster

production heap profiling for the JVM. compatible with google-perftools.
https://github.com/mariusae/heapster

Last synced: 10 months ago
JSON representation

production heap profiling for the JVM. compatible with google-perftools.

• Host: GitHub
• URL: https://github.com/mariusae/heapster
• Owner: mariusae
• Created: 2010-08-28T00:59:02.000Z (over 15 years ago)
• Default Branch: master
• Last Pushed: 2016-05-11T21:12:18.000Z (over 9 years ago)
• Last Synced: 2025-03-31T10:09:31.062Z (10 months ago)
• Language: C
• Homepage:
• Size: 81.1 KB
• Stars: 395
• Watchers: 27
• Forks: 21
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-jvm - heapster - Production heap profiling for the JVM. (Profilers)

README

          
[![Build Status](https://secure.travis-ci.org/mariusae/heapster.png)](http://travis-ci.org/mariusae/heapster)

# Heapster

Heapster provides an agent library to do heap profiling for JVM

processes with output compatible with

[Google perftools](https://github.com/gperftools/gperftools). The

goal of Heapster is to be able to do meaningful (sampled) heap

profiling in a production setting.

Currently it allows for profiling similar to the TCMalloc library,

e.g.:

    $ HEAPSTER_PROFILE=/tmp/OUT java -agentlib:heapster Test

    $ pprof /tmp/OUT

    Welcome to pprof!  For help, type 'help'.

    (pprof) top

    Total: 2001520 samples

     2000024  99.9%  99.9%  2000048  99.9% LTest;main

        1056   0.1% 100.0%     1056   0.1% Ljava/lang/Object;

         296   0.0% 100.0%      296   0.0% Ljava/lang/String;toCharArray

         104   0.0% 100.0%      136   0.0% Ljava/lang/Shutdown;

By default, Heapster samples every 512 kB, this can be changed with

the environment variable `HEAPSTER_SAMPLE_PERIOD` (in bytes).

This is still work in progress.

# Installation (Example)

    $ make

    $ cp libheapster.dylib /usr/local/lib/

# Twitter Server Integration

If you use [Twitter Server](https://github.com/twitter/twitter-server), and run your

program with heapster, you can generate runtime heap profiles like so:

    $ curl 'localhost:9990/admin/pprof/heap?pause=10&sample_period=1024' > /tmp/prof

This will collect heap growth for 10 seconds, with a sampling period

of 1kB.